1
$\begingroup$

Suppose we have Heisenberg equation of motion for some observable $A$,

$$ i\hbar\frac{dA}{dt}= -[H,A] $$ since the trace of any finite dimensional commutator structure vanish(not something like $[x,p]=i\hbar$ ),

$$ Tr\left(\frac{dA}{dt}\right)=\frac{i}\hbar~Tr([H,A])=0 $$

so my question is, what does $Tr(\frac{dA}{dt})=0$ indicate?

Improve this question
$\endgroup$
2
  • $\begingroup$ $A$ and $H$ are typically not finite dimensional, the energy spectrum is typically infinite, so you cannot use commutativity of Trace. $\endgroup$
    – user7757
    Commented Jun 12, 2014 at 19:22
  • $\begingroup$ @ramanujan_dirac I am not talking about [x,p]=ih, in many cases people are using heisenberg equation of motion, and applied to finite dimensional system. $\endgroup$
    – Lorniper
    Commented Jun 12, 2014 at 19:27

1 Answer 1

Reset to default
1
$\begingroup$

If the trace of an operator with a complete set of eigenvectors is well defined, then it is equal to the sum of the eigenvalues of the operator.

Time evolution of an operator is a unitary transformation, which leave the eigenvalues of an operator unchanged. This means if $a_i$ is an eigenvalue of an operator $A$

\begin{equation}\frac{\mathrm{d}a_i}{\mathrm{d}t} = 0\end{equation}

and so clearly

\begin{equation}\mathrm{Tr}\left(\frac{\mathrm{d}A}{\mathrm{d}t}\right) = \sum_i\frac{\mathrm{d}a_i}{\mathrm{d}t} = 0\end{equation}

So physically this result is to do with the fact that although the system may evolve with time, the spectrum of allowed results for a measurement of $A$ does not.

Improve this answer
$\endgroup$
2
  • $\begingroup$ "Time evolution of an operator is a unitary transformation", I don't think this's always true, right? $\endgroup$
    – Lorniper
    Commented Jun 13, 2014 at 5:38
  • $\begingroup$ We can write $A = U^\dagger A_0U$, where $A_0$ is time independent. The solution to the Heisenberg equation for $U$ is then $U = \exp\left(-\imath\frac{Ht}{\hbar}\right)$ which can be shown to be unitary for Hermitian $H$ $\endgroup$
    – By Symmetry
    Commented Jun 13, 2014 at 9:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.